Retrospective Evaluation of Ultrasound Features of Thyroid Nodules to Assess Malignancy Risk: A Step Toward TIRADS

OBJECTIVE. The aim of this retrospective study was to develop a thyroid nodule scoring system for malignancy potential to better select nodules for ultrasound (US)-guided fine needle aspiration (FNA). MATERIALS AND METHODS. US-guided FNA was performed on 2375 thyroid nodules in successive patients. Cytologic or histopathologic confirmation of disease state in 2002 lesions showed that 148 were malignant. We developed an extended scoring system to provide more decision levels than standard scoring by including weak (macrocalcification, eggshell calcification, hypoechogenicity, solid consistency) as well as strong indicators of malignancy (microcalcification, hypoechogenicity, lobulated or ill-defined margins, taller-than-wide shape, suspicious lymph nodes) and by including contraindications (hyperechogenicity, comet-tail artifact, complete halo, cystic or microcystic). ROC analysis was used to compare detection accuracy and decision thresholds resulting from standard scoring using five major features with extended scoring. RESULTS. Although an accuracy advantage was found for the extended scoring over standard scoring, we discount this finding because our scoring involved a preliminary analysis. However, the extended scoring offers more reporting options for certainty of malignancy: standard scoring gave four potential thresholds for reporting; extended scoring gave nine. The most useful of the additional thresholds captured nearly 88% of malignancies but resulted in only 28% of patients without malignancy undergoing biopsy. CONCLUSION. Our extended Thyroid Imaging Reporting and Data System scoring provides more operating points to support treatment decisions. A simplified decision rule-biopsy every nodule with at least two weak features or one strong feature-preserves the most useful of the new decision thresholds from extended scoring.